search

aws-samples / aws-eda-slurm-cluster

AWS Slurm Cluster for EDA Workloads
MIT No Attribution
28 stars 7 forks source link

[BUG] ParallelCluster database stack should include slurmdbd instance and slurmdbd #147

Closed cartalla closed 3 weeks ago

cartalla commented 1 year ago

Describe the bug Currently ParallelCluster provides a CloudFormation template to create the Slurm database. This stack only creates an RDS database, but doesn't create an instance to run the slurmdbd daemon. This doesn't follow the recommended architecture and limits the ability of multiple cluster to share the same accounting database. Currently ParallelCluster runs a slurmdbd daemon on the controller of each cluster that is using the database. This isn't the recommended architecture and I'm not sure if this will work.

cartalla commented 3 weeks ago

As of ParallelCluster 3.10.0, ParallelCluster now supports an external slurmdbd instance and support for this has been added to this project.